The Testing in GEE Method without PA Condition

Marginal regression model and its associated generalized estimating equation (GEE) are becoming increasingly being used in longitudinal studies. Pepe and Anderson (1994) pointed out that there is an important assumption called PA condition behind GEE method. If the assumption is violated and nondiagonal working correlation matrix is used in GEE, the statistical inference may be deficient. This paper focused on PA condition's influence on the testing of regression coefficients in GEE method by theoretic and numeric analysis. Due to the violation of the PA condition, the distributions of Wald statistics and Score statistics based on GLS estimators are noncentral \chi^2 distributions. The efficiency of testing based on the GEE method is largely influenced.